Hello all, we are currently evaluating a migration from Xen to KVM.
We host our VMs on a SAN, access to LUNs is done by the iSCSI protocol, all multipathed.
With Xen, we faced a problem when VMs were consuming a lot of CPU and a lot of I/O. In this situation, the Dom0 did not get enough CPU time to process the I/O requests through the iSCSI processes and to run the iscsid daemon. So, the Dom0 was always loosing the iSCSI connection to the SAN because of timeout (node.conn[0].timeo.noop_out_timeout).
We solved it by dedicating a core to the Dom0, as recommended in the Xen FAQ.
Can this happen with KVM ? Why can(not) it happen ? If it could happen, what is the recommended solution (renice, cgroups, ...) ?
Thanks for your help. -- Romain LE DISEZ -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html